Having spent the last two articles explaining why chemotherapy as currently practiced is a terrible way to treat metastatic cancer (if the goal is cure or even just life extension) it’s time to pivot a bit and explain that while standard chemotherapy “best practices” are completely the wrong approach for the majority of cancer patients, chemo drugs can still have some valuable utility, especially as part of a multifaceted treatment, if doctors applied scientific knowledge from other fields to their understanding of why even effective chemo drugs nearly always become eventually useless in the same patient.
While any oncologist is familiar with the problem of acquired resistance to previously effective chemo drugs, it seems that the field of clinical oncology as a whole has failed both to examine carefully the mechanisms for this or to explore possible methods to overcome or delay resistance (other than by the ultimately useless blunt instrument of switching drugs), in order to extend meaningful life in metastatic patients.
There are two important facets to this failure. The first, briefly, is cultural. Our society’s Cancer Machine has an odd, collective cognitive dissonance in which the idea of “curing” late-stage patients, understood rationally to be nearly impossible, exists alongside treatment techniques theoretically designed to cure, but which never have come close to this ideal. (This dovetails with the other cognitively dissonant dichotomy in which the continued propaganda of “cure” overlays the fear of cancer as inevitable death in early-stage patients.) So doctors, particularly specialists such as surgeons and radiologists treating one part of the body, often imply that the goal of a given treatment is cure, even while their own data show that cure is essentially never achieved by their treatments.
But the focus of the discussion here is the science of tumors: their heterogeneity and ability to evolve, which oncologists again understand in the abstract, but generally fail to respond to, because they do not seem to understand evolutionary biology. This is not for lack of researchers encouraging the approach to tumors in these terms (Gerlinger and Swanton 2012; Greaves and Maley 2012; Marusyk et al 2012). And it’s not doctors’ fault that medical schools that their training does not include the basic principles of evolution and population biology (although in the case of oncologists at least, it clearly should).
Here is an oversimplification of the essential problem (for more nuances and explanations showing how much more complicated it actually is, see the references below): malignant tumors are heterogeneous populations of cells which actually compete with each other for resources. Tumor cells are mutating all the time, via several mechanisms such as drift and varying selection pressures, which include exposure to different chemical microenvironments, oxygen levels, etc. (which can even be different in different parts of the same tumor, not just between different tumors).
Although there are many selective pressures driving tumor evolution that we have no known control over, the use of chemotherapy is a big one that we do. That is, our misuse of chemotherapy for cancer treatment is similar to the misuse of pesticides in agriculture. In the latter case we create resistant pests by imposing a strong selection pressure on a heterogeneous population of insects, most of which are susceptible to the poison but a few of which by chance are resistant. Once all the susceptible individuals are destroyed, the resistant ones have no competition and can take over, and the pesticide is now useless.
In a way, chemotherapy is even more insidious because of the evidence that chemo itself increases the mutation rate of cancer cells, which is abnormally high. The implications of this should be sobering: misuse of chemo drugs can actually hasten the evolution of one’s tumors to a more aggressive form (although it has been argued that the changes occur more on a phenotypic than genotypic level [Germain, 2012]). In any case, the end result is that the current paradigm of chemotherapy treatment, the delivery of “highest tolerated dose” (HTD), while intended to circumvent resistance by the use of overwhelming force (as the use of high levels of antibiotics for ten days to two weeks wipes out an infection so that it will not return), in reality almost never works, mainly because the levels of drugs truly needed to wipe out all populations of cancer cells would be much too high for the patient to survive. So, we continue to deliver drugs at a sublethal level which promotes resistance, and probably shortens lifespan much more often than currently acknowledged. The effects of even those drugs identified in trials as providing “complete response” for some patients are nearly always only temporary, and extend life in many cases by less than a year, by which time resistant (untreatable) tumors come roaring back (the dreaded “progressive disease”), ultimately killing the patient for whom there are no more treatment options.
There is already a fairly rich literature on the mechanisms of the evolution of resistance in tumors, although ideas to overcome it are still largely theoretical and remain untested. One study, however, provides not only a model for how resistance could possibly be overcome, but tests the model successfully in mice in a laboratory setting, with a quite intriguing result. Gatenby et al. (2009) developed a mathematical model that showed that it should be possible to maintain a low tumor burden with a single effective chemo drug, by continually adjusting the rate and levels (and thus the selective pressure on the tumor cell population) at which the drug is delivered. The key to the new paradigm is that it does not aim for a cure, but rather management of tumors at a non-life-threatening level. To shift to such a paradigm would not be unreasonable scientifically, since cure of metastatic disease via conventional treatments is essentially nonexistent (no matter how much we pretend it is not). The authors find that lifespan can be extended indefinitely by using this protocol that is individualized to how the patient’s tumors are responding to the drug.
The important assumption of the model is that resistance implies a metabolic cost, which allows susceptible cells to outcompete resistant cells in the absence of the selective chemo agent. While this may not necessarily be true for all forms of resistance (Germain, 2012) there is a lot of evidence that it is much of the time (including from Gatenby et al.’s subsequent laboratory experiment). Given that assumption, chemotherapy treatments are adjusted downward when there is a strong response from a tumor — i.e., when the tumor shrinks quickly as the majority population of susceptible cells is killed. This allows susceptible cells to remain and continue to compete with resistant cells, keeping them suppressed to a relatively small population size.
What strengthens Gatenby et al.’s paper is that they followed up their mathematical model with an in vivo mouse study, testing the model’s predictions. Human tumors implanted in mice are convenient for this purpose because they can be measured easily (see picture). In the authors’ treatment group, tumors were measured and the chemotherapy dose adjusted, while for comparison a group of mice was administered the typical static dose of the same chemo drug, regardless of tumor size. Not only was a low tumor burden maintained in the dynamically treated mice (which resulted in a greatly increased lifespan), but the amount of the drug required to maintain the tumors at a low level dropped to 17% of the normally administered levels of the drug, suggesting that much lower toxicities are needed under this protocol. In other words, it’s a win-win: longer life at a higher quality. The only thing that is being given up in return is the misguided hope for a cure that does not exist anyway.
The time when cancer for many more people can be managed as a chronic disease is thus possible with the tools we already have. This pre-clinical result needs to be followed up with human data, in order to establish how to make such a protocol work in a situation where the tumors cannot be so easily measured with a pair of calipers. Refining this will clearly take time, given all the permutations of cancers and drugs, but now is the time to start, and no doubt there are many stage IV patients who would be willing to try something different from the hopeless alternative. What is clear is that for all the reasons stated here and in the last post, standard chemotherapy regimens for metastatic patients are clearly not only a waste of time, but worse, are killing a lot of us more quickly and miserably than if we just died of our cancer.
There’s no need to throw out the chemo baby with the bathwater, as shown by this paper; we just need to be a lot more intelligent with its use, and that will start by cancer doctors listening to what evolutionary biologists have to offer in their understanding of tumor biology.
Gatenby RA, Silva AS, Gillies RJ, Frieden BR, 2009. Adaptive therapy. Cancer Res. 69(11):4894-903. doi: 10.1158/0008-5472.CAN-08-3658.
Gerlinger M, Swanton C 2012. How Darwinian models inform therapeutic failure initiated by clonal heterogeneity in cancer medicine. Br J Cancer. 103(8):1139-43. doi: 10.1038/sj.bjc.6605912. Epub 2010 Sep 28.
Germain, P-L., 2012. Cancer cells and adaptive explanations. Biology & Philosophy 27(6):785-810.
Greaves M, Maley CC 2012. Clonal evolution in cancer. Nature. 481(7381):306-13. doi: 10.1038/nature10762.
Marusyk A, Almendro V, Polyak K 2012. Intra-tumour heterogeneity: a looking glass for cancer? Nat Rev Cancer. 12(5):323-34. doi: 10.1038/nrc3261.